ProjectSummary Mostlaboratoriesstudyingbiologicalprocessesandhumandiseaseusemicroscopestoimagesamples. Whetherinsmallorlargescalemicroscopyexperiments,biologistsincreasinglyneedsoftwaretoidentifyand measurecellsandotherbiologicalentitiesinimages,toimprovespeed,objectivity,and/orstatisticalpower. Theprincipalinvestigatorenvisionsbringingtransformativeimageanalysisandmachinelearningalgorithms andsoftwaretoawideswathofbiomedicalresearchers.Inadecade,researcherswilltacklefundamentally newproblemswithquantitativeimageanalysis,usingseamlessimagingworkflowsthathavedramaticnew capabilitiesgoingbeyondtheconstraintsofhumanvision. Tothisend,thePIwillcollaboratewithbiologistsonimportantquantitativeimagingprojectsthatalsoyield majoradvancementstotheiropensourceimageanalysissoftware,CellProfiler.Thisversatile,userfriendly softwareisindispensableforbiomedicalresearch.Launched125,000+times/yearworldwide,itiscitedin 3,400+papersfrom1,000+laboratories,impactingahugevarietyofbiomedicalfieldsviaassaysfromcounting cellstoscoringcomplexphenotypesbymachinelearning.CellProfilerevolvesinanintenselycollaborativeand interdisciplinaryresearchenvironmentthathasyieldeddozensofdiscoveriesandseveralpotentialdrugs. Still,manybiologistsaremissingoutonthequantitativebioimagingrevolutionduetolackofeffective algorithmsandusablesoftwarefortheirneeds.InadditiontomaintainingandsupportingCellProfiler,theteam willimplementbiologistrequestedfeatures,algorithms,andinteroperabilitytocopewiththechangingland scapeofmicroscopyexperiments.Challengesincludeincreasesinscale(sometimesmillionsofimages),size (20+GBimages),anddimensionality(timelapse,threedimensional,multispectral).Researchersalsoneedto accommodateavarietyofmodalities(superresolution,singlemolecule,andothers)andintegrateimage analysisintocomplexworkflowswithothersoftwareformicroscopecontrol,cloudcomputing,anddatamining. ThePIwillalsopioneernovelalgorithmsandapproacheschangingthewayimagesareusedinbiology, including:(1)afundamentalredesignoftheimageprocessingworkflowforbiologists,leveragingrevolutionary advancementsindeeplearning,(2)imageanalysisformorephysiologicallyrelevantsystems,suchasmodel organisms,humantissuesamples,andpatientderivedcultures,and(3)datavisualizationandinterpretation softwareforhighdimensionalsinglecellmorphologicalprofiling.Inprofiling,subtlepatternsofmorphological changesincellsaredetectedtoidentifycausesandtreatmentsforvariousdiseases.Wewillalso(4)integrate multipleprofilingdatatypes:morphologywithgeneexpression,epigenetics,andproteomics.Ultimately,we aimtomakeperturbationsincellmorphologyascomputableasotherlargescalefunctionalgenomicsdata. Overall,thelaboratory?sresearchwillyieldhighimpactdiscoveriesfrommicroscopyimages,andits softwarewillenablehundredsofotherNIHfundedlaboratoriestodothesame,acrossallbiologicaldisciplines.